My company operates several robot welding stations across multiple sites, and our IT team supports a fair amount of operational technology without having a dedicated OT specialist. The controllers currently use USB drives for backups, but the equipment owners want an automated process.
I'm considering placing the Ethernet-connected controllers on a dedicated OT VLAN with appropriate network restrictions, then backing up their TP program files monthly. We normally use Veeam for business systems, but these controllers vary in age and capability. Some may only support older protocols, and installing backup software directly on a machine that controls production could introduce compatibility or reliability risks.
What backup architecture have others used for industrial robot controllers? Would a secured jump box, vendor-supported file transfer, Veeam, or a simple scheduled script be the best approach? I'd also appreciate advice on backup frequency, validation, and whether full disk images are more appropriate for older or especially fragile systems.
4 Answers
Don’t assume that a more sophisticated backup product is automatically better for production equipment. A small, well-tested script launched by Task Scheduler can be safer and have less impact than installing a persistent third-party agent on a robot controller. Copy the files to a staging or jump system, retain multiple dated versions, and have a separate process verify that every expected controller completed its backup. Weekly or daily backups may be more appropriate than monthly if the programs change frequently.
For controllers with unusual operating systems, unsupported software, or no reliable secure file-transfer option, offline imaging may be the most dependable recovery plan. A tested clone of the controller’s drive can get a replacement system running quickly, while another copy is stored separately. Keep at least one disconnected or otherwise protected copy, document the restoration procedure, and test that the backup can actually be restored rather than only checking that a file was created.
If the controllers are actually Windows systems and are modern enough to support the required protocols, check whether Veeam can protect them directly. Confirm this with the equipment vendor first, especially around SMB support and whether any backup agent is approved. Older controllers may only support SMBv1 or other insecure legacy methods, so putting them behind a restricted jump host is usually safer than enabling old protocols throughout the network.
A common pattern is to use a hardened jump box that has tightly controlled access to the OT network. Have the controllers send or copy their program files to that system, then let your normal backup platform protect the jump box. That keeps the backup software and broader management traffic away from the controllers themselves. Make sure the firewall rules are limited to only the required hosts and ports, and monitor the job so a failed copy does not go unnoticed.
I’d verify the controller model and supported SMB versions with the vendor before designing around Veeam. If SMBv2 or SMBv3 is available, the setup should be much easier; if the device only supports SMBv1, isolate that communication as narrowly as possible rather than enabling it broadly.